- Sample Space: Ω={(r1,r2,…),ri∈{0,1},∀i≥1} - Event Space: For any n≥1 and a1,…,an∈{0,1} set Aa1,…,an={(r1,r2,…)∈Ω:ri=ai,1≤i≤n}and then let J:={Aa1,…,an:n≥1,a1,…,an∈{0,1}}∪{∅,Ω}.We want (at minimum) a σ-algebraF⊇J with: - Probability Measure: P(A∈Fa1,…,an)=2n1,∀n≥1,a1,…,an∈{0,1}
\begin{proof} 1. ∅,Ω∈J 2. Closed under intersection since each event is either disjoint (i.e. its made up of discrete permutation of events) or they’re the same meaning the intersection is the same. 3. Complement=finite union: 1. Use idea that (0,0)c=(0,1)∪(1,1)∪(1,0) and then apply this at level n. \end{proof}
There is a probability space(Ω,F,P∗) such that 1. F⊇J and; 2. P∗(A)=P(A) for A∈J
\begin{proof} 1: Use Extension Theorem, need to show P is countably additive on J. Clearly, P is finitely additive: P(Aa1,a2∪Aa1,a2c)=P(Aa1)=21P(Aa1,a2)+P(Aa1,a2c)=221+221=21 Countably requires Compactness argument (see lemma 2.62 in book) 2: Compare Ω to (0,1] via binary expansion. \end{proof} >[!lemma] >Every x∈(0,1] has a unique binary expansion x=j>1∑2jrjsuch that rj∈{0,1} and rj=1i.o.. >